Wound Contraction in Relation to Collagen
Formation in Scorbutic Guinea-pigs
by M. ABERCROMBIE, M. H. FLINT, and D. W. JAMES 1
From the Department of Anatomy and Embryology, University College London
INTRODUCTION
A W O U N D in any mobile part of the skin of a mammal diminishes in area as it
heals by a centripetal movement of the undamaged skin surrounding it. This
movement, usually called wound contraction, depends on a pull exerted by the
material within the wound (Lindquist, 1946; Abercrombie, Flint, & James, 1954;
Billingham & Medawar, 1955). It is commonly believed that the effective force is
developed by the newly formed collagen fibres. In a previous paper, however
(Abercrombie, Flint, & James, 1954), we found that the course of the contraction
of skin wounds in rats did not parallel the deposition of new collagen, chemically
measured. This result, while certainly in no way conclusive by itself, suggested
that the supposed role of collagen in contraction ought to be tested more
stringently. This we have now done by measuring wounds made on guinea-pigs
receiving a diet devoid of ascorbic acid. Such a deficiency largely prevents the
formation of new collagen (reviewed by Wolbach & Bessey, 1942). We found,
however, that it did not prevent wound contraction, a result difficult to reconcile with the usual hypothesis. We have accordingly put forward a new hypothesis as to the causal agent of contraction, which we suggest is a force produced
by the population of connective tissue cells within the wound.
MATERIAL AND METHOD
In the main experiment thirty adult male guinea-pigs of mixed stock were
used; for convenience of tattooing white or mainly white animals were selected.
Their mean body-weight at the beginning of the experiment was 442 g. (range
378-570, standard deviation 43-4). They were tattooed under ether anaesthesia
(Abercrombie, Flint, & James, 1954), eight needles being operated simultaneously to mark a square of an area, measured through the centres of the eight
tattoo points, of a mean size of 255 mm.2 (range 181-34-4, standard deviation
1
Authors' address: Department of Anatomy & Embryology, University College London,
Gower Street, W.C.I, U.K.
[J. Embryol. exp. Morph. Vol. 4, Part 2, pp. 167-75, June 19561
168
M. A B E R C R O M B I E , M. H. F L I N T , A N D D. W. J A M E S
4-11). The marks were made in the dorsilumbar region, one square on each side
of the mid-line. After tattooing the animals were put on the ascorbic acid deficient diet of Murray & Kodicek (1949) with the salt mixture of Hubbell, Mendel,
& Wakeman (1937), but half of them (chosen by random numbers) received in
addition a supplement of crystalline ascorbic acid mixed, in the proportion of
1 g. to 100 g., into their food. After 5 days on the diet aseptic operation was performed on each animal under ether anaesthesia. A tracing of the tattoo marks
was first taken on tracing paper, and a square wound of about 17 mm.2 was made
within the tattoo points of one of the pair of squares marked on each animal, the
second square remaining for estimation of changes in normal skin during the
experiment. The wound extended through the panniculus carnosus to the deep
fascia. No dressing was applied. Ten days later the animals were killed. The
position of the tattoo marks was again marked on tracing paper, the scab
removed (it was necessary to remove the larger scabs before tracing), and the
new-formed tissue within the tattoo marks excised. A few animals were rejected
on suspicion of infection and we were finally left with material from 14 deficient
and 16 non-deficient animals.
The size of each tattoo area, marked by the centres of the tattoo points as
described in our previous paper, was obtained by pricking through the tracings
on to black paper, cutting out the area enclosed and measuring the transmission
of light through the resulting hole photoelectrically, comparing standards of
known area. The collagen content of the scabs and that of the excised tissue were
obtained separately by the method of Neuman & Logan (1950), which involves
estimating hydroxyproline. The results are recorded as milligrammes of hydroxy-
proline. They may be converted to milligrammes of collagen by multiplying
by 7-46.
Mean body-weight on the first day of the diet was 429 + 9 g. in the deficient
group, 453 +12 g. in the non-deficient group. The difference is not significant
(t = 1-50,28 degrees of freedom, 0 2 > P > 0-1). Weight-loss during the 15 days
of the experiment was practically the same in the two groups (the basal diet was
not well taken): in the deficient group it was 46 + 7 g., in the non-deficient group
45 + 7 g. The initial tattoo area within which the wound was made measured
2 5 3 + 11 mm.2 in the deficients, 256 + 1 0 mm.2 in the non-deficients. In a
sample of 10 of the deficient group and 8 of the non-deficient group the wet
weight and hydroxyproline content of the piece excised at the initial wounding
were obtained, to check that similar wounds had been made. The mean wet
weights of these pieces were a little different though not significantly so (deficients 25-4+1-8 mg., non-deficients 21-3 + 13 mg.; t= 176,16 d.f.,P=01). The
excised tissue of the deficients had a mean hydroxyproline content of 633 + 45 /xg.,
that of the non-deficients 577 + 33 /xg., again a non-significant difference (t = 0-96,
]6d.f.,0-4>P>0-3).
WOUND CONTRACTION AND COLLAGEN
169
RESULTS
Hydroxyproline of repair tissue
At autopsy, after careful removal of the scab, the tissue within the border of
the wound marked by the tattoo was excised. It was markedly different in
appearance and consistency in the deficient and non-deficient groups. In the
deficient group it was very friable, and in the non-deficient control group it was
quite firm. Its wet weight was similar in the two groups, with a mean of 13 0 +
17 mg. in the deficient and 10-3 ± 1 3 mg. in the control group. Its hydroxyproline content was, on the contrary, highly dissimilar, with a mean of 104 +
2 6 ng. in the deficient and 74 + 8 /xg. in the control group. It is clear that the
vitamin deficiency was highly effective in reducing collagen formation. The small
amount of hydroxyproline detected in most of the deficient animals (in 3, none
was detectable) may, however, represent a trace of reticulin and collagen formed.
Silver-impregnated fibres were found in a sample wound fixed in Bouin and
stained with Wilder's reticulin method; but no evidence of collagen appeared
with Mallory's connective tissue stain. The formation of a small quantity of
argyrophilic reticulin (e.g. Bourne, 1944; Penney & Balfour, 1949) and even of
mature acidophilic collagen (e.g. Hartzell & Stone, 1942; Danielli, Fell, &
Kodicek, 1945) has been described in sections of wounds from scorbutic animals.
Wolbach (1933) considered that such fibres do not develop in a totally vitamin C
depleted guinea-pig. Since our animals were on the deficient diet for only 5 days
before operation it is possible that their reserves of ascorbic acid were not entirely exhausted before healing. There is, however, the possibility to be considered that the hydroxyproline found was not present in the form of collagen
but as some soluble precursor or break-down product. We tried but failed to
obtain evidence in support of this possibility. In half the specimens the estimation was made of total hydroxyproline regardless of its solubility, and in half the
estimation was made after extraction in 20 per cent, urea so that it represented
that part of the hydroxyproline having the solubility of collagen. In the deficient
group the total hydroxyproline was 13 + 6 {N= 8), the collagen hydroxyproline
7-3 + 2-4 (N = 6). In the control group the total was 71 + 11 (N = 7), the collagen
76 +12 (N = 9). The differences within each group between total and collagen are
obviously not significant. For this reason the separate types of estimation have
been combined.
Scabs
There was usually a striking difference between the deficient and control
animals in the scab covering the wound 10 days after operation. It was usually
small both in area and thickness in the controls, averaging 3-7 + 1-4 mg. wet
weight (N= 16), including three where it had disappeared entirely. In the deficients it was usually very much thicker and larger in area. In this group it averaged 28 + 6 mg. wet weight (N= 14), including two where it had disappeared
entirely. Some of the scabs of control animals contained traces of hydroxyproline,
170
M. A B E R C R O M B I E , M. H. F L I N T , A N D D. W. J A M E S
as found also at 10 days after wounding in rats (James, 1955): the average for
those scabs still present was 10 + 8 /xg. (N=13). Most of the scabs from the
scorbutic animals, on the other hand, contained substantial amounts of hydroxyproline: their mean was 112 + 23 /xg. (N = 12). It is unfortunately impossible to
determine reliably from these data whether the concentration of scab hydroxyproline was significantly higher in the deficient group or not, since only wet
weights are available and water content is not likely to be comparable; and in
any case the variances of the two groups are too disparate for proper comparison.
The total hydroxyproline associated with the wound, that is to say both in the
scab and in the underlying repair tissue, does not differ significantly between the
deficient and control groups. Their means are respectively 121 +23 (N= 14), and
82 +10 (N= 16), and comparing these f = l - 6 5 , 0 - 2 > P > 0 - l . This might raise
the suspicion that in the deficient animals the hydroxyproline-containing part
of the repair tissue was torn away with the large and adherent scab, which would
make spurious the difference in hydroxyproline content found between the
repair tissue of the deficients and controls. Some damage to the repair tissue did
in fact occasionally appear to occur in this way; but the inclusion of an important
amount of the hydroxyproline content of the wound with that of the scab should
produce a strong negative correlation between scab and wound hydroxyproline,
of which there was no trace: the correlation coefficient was positive but nonsignificant ( + 010).
Contraction
Our previous investigation (1954) of skin wounds in rats showed that the
difference between the area marked out by the centres of the tattoo points at
operation and the equivalent area at autopsy gives a close approximation to
(about a 20 per cent, exaggeration of) the contraction of the actual wound within
the tattoo points. Measured through the tattoo points, mean contraction in the
deficient group during the 10 days of healing was 8 3 + 1 0 mm.2 (N= 14); in the
non-deficient group it was 1 0 1 + 14 mm.2 (AT =16). Significant contraction
obviously occurs in the vitamin C deficient group (^=829, P < 0001); and the
amount that occurs is not significantly different from that of the non-deficient
control group (/ = 102, 0 4 > P > 03). In three of the deficient group with no
detectable hydroxyproline, contraction was 6 8, 9 4, and 12-5 mm.2 Qualitatively, the contraction appeared to be of the same nature, in that the tension was
released when the content of the wound was cut free from the margin of normal
skin. Our data therefore fail to provide any evidence that contraction is dependent on collagen formation. Nevertheless, the small difference actually found was
in favour of the control group; and when both groups are pooled, the six wounds
with most contraction all belong to the non-deficient group. It might be suspected
that with larger samples the slight difference would become significant. This may
be so, but the data do not suggest that such a difference could then be connected
with the hydroxyproline content of the repair tissue. In fact in both deficient and
WOUND CONTRACTION AND COLLAGEN
171
control groups wound hydroxyproline and contraction were negatively, though
not significantly, correlated (r= - 0 2 6 in both groups). The scab might have
been expected to obstruct contraction in the deficient group; and scab weight
was indeed negatively correlated with contraction in this group (r= - 035) but
again the relation is not significant at the 5 per cent, level of probability.
Changes in normal skin
There was a small and non-significant diminution of the area of the tattoo on
the control side during the 15 days between the start of the diet and autopsy in
a sample of both deficient and non-deficient groups (diminution of 0 7 + 2-2mm.2,
N= 8; and 0 8 + 0 7 mm.2, N=6 respectively). The possibility that the deficient
diet might produce detectable loss of collagen was also investigated in a small
sample. No significant loss of collagen could, however, be detected when we
compared the content of the piece removed at operation with the content of the
control area which was removed at autopsy. At operation the mean content was
608 + 33 /*g. (N = 9), at autopsy 700 + 70 /xg. (N = 9). At autopsy the weight of the
piece removed was rather higher than at initial operation, though not significantly so; probably this was the result of a slight bias in choosing the best-marked
side for operation. Use of analysis of covariance to eliminate the size difference
left the hydroxyproline difference still quite non-significant (?=0-383, P = 07).
DISCUSSION
The inhibition of collagen formation in healing wounds by ascorbic acid
deficiency has been demonstrated many times before (see Wolbach & Bessey,
1942), though apparently not previously by a chemical method. The absence of
any effect of the deficiency on the amount of already formed collagen has been
shown chemically (Elster, 1950; Robertson, 1950, 1952). The conspicuous and
persistent scabs that form over wounds of scorbutic guinea-pigs were noted by
Wolbach & Howe (1926), Hartzell & Stone (1942), and Danielli, Fell, & Kodicek
(1945). It is not clear why they should be so different from the scabs of normal
animals, but the increased fragility of the vessels (Lee & Lee, 1947) may mean a
greater production of exudate; and delayed epithelialization (Danielli, Fell, &
Kodicek, 1945; Galloway, Garry, & Hitchin, 1948) may allow this to accumulate.
We have, however, no reason to believe that epithelialization was delayed in our
specimens: the presence of the scab does not necessarily mean a failure of healing, since one specimen with a particularly large scab, which was examined
histologically, had a complete and hyperkeratinized epidermis adherent to the
underside of the scab. Because of the absence of collagen, such epidermis is easily
torn away when the scab is removed, and the wound then looks unhealed. The
presence of hydroxyproline in the scabs of wounds in rat-skin has already been
demonstrated by James (1955). It remains uncertain whether scab hydroxyproline represents a diffusible potential precursor of new collagen, or comes from
degeneration of some of the original collagen bordering the wound. Devenyi &
172
M. A B E R C R O M B I E , M. H. F L I N T , A N D D. W. J A M E S
Holczinger (1954) have described the incorporation of degenerating connective
tissue of the wound floor into the scab. It is not at present worth speculating on
the difference in scab hydroxyproline between control and scorbutic animals. It
may merely be a reflection of the curious persistence of the scab in the scorbutic
animals, and not of any difference in hydroxyproline production.
The main purpose of this work was to investigate the relation between wound
contraction and collagen formation. We found that wound contraction was not
significantly reduced by the very severe inhibition of collagen formation that
resulted from the ascorbic acid deficiency. On the average only 15 per cent, of
the amount of hydroxyproline in the wounds of the controls was present in the
wounds of the scorbutics, and three of the latter contracted even though devoid
of detectable hydroxyproline. The contraction in both groups involved a diminution of the actual wound area by probably about 30-40 per cent., a considerable
proportion, though rather less than occurs during the first 10 days in the healing
of a similar-sized wound in the rat, where it is about 60 per cent. (Abercrombie,
Flint, & James, 1954). The traditional hypothesis implicating newly formed
collagen will not reasonably account for this contraction. Nor does there seem
to be any good reason for suggesting that some intercellular component other
than collagen is responsible. An intercellular substance is formed in scorbutic
wounds. It has been described by Wolbach (1933), Penney & Balfour (1949),
and Bradfield & Kodicek (1951). It is, however, histologically and histochemically quite different from that of normal wounds. It is appropriate therefore to
suggest a new hypothesis, bearing in mind that perhaps the whole range of
phenomena included under wound contraction in man and other species may
require more than one explanation. The new hypothesis is that the contractile
force is produced by the connective tissue cells that occupy the wound. Many
authors have remarked that cellular proliferation in wounds is undiminished, or
even slightly increased, by ascorbic acid deficiency (Wolbach, 1933; Hunt, 1941;
Hartzell & Stone, 1942; Bourne, 1944; Meyer & Meyer, 1944). The fibroblasts
may not be normal in form (Penney & Balfour, 1949) or in arrangement (Danielli,
Fell, & Kodicek, 1945; Meyer & Meyer, 1944), and their immigration may be
delayed (Mazoue, 1937). But, at least as far as microscopic observation extends,
they are much less affected than is the intercellular substance. This is in contrast
with wounds in animals treated with cortisone. Here both collagen formation
and the recruitment of new cells is depressed. It is therefore not surprising, from
the standpoint of the hypothesis we are putting forward, that contraction is inhibited too (Billingham, Krohn, & Medawar, 1951 a, b).
In one respect cells are a more obvious first choice for a contractile mechanism
than is collagen, since neither a reticulin meshwork nor individual collagen
fibres have ever been shown to be contractile under physiological conditions, but
contractile mechanisms obviously occur at least in some cells. Smooth musclecells in tissue culture are indistinguishable from ordinary connective tissue
fibroblasts; and it is conceivable that the contractile power of muscle-cells is
WOUND C O N T R A C T I O N AND COLLAGEN
173
widespread in a primitive form throughout the fibroblast family (see HoffmannBerling, 1954). This line of thought, however, perhaps directs attention too much
to a contractile mechanism intrinsic to the cell. Another possibility is that the
force may be produced by the mutual rearrangement of cells, perhaps by expansion of the areas of adhesion which exist between cells in a fibroblast colony
(Kredel, 1927), so packing them more closely together. Whatever the mechanism,
there is no doubt that cells that are not muscle-cells can exert a tractive force.
Mayer (1933) has shown that considerable tensions exist in fibroblast cultures
which can be ascribed to the cells. Twitty (1949) has found that melanoblasts of
Triturus under certain circumstances draw themselves together, and this is
important in the formation of pigment patterns. Finally, a possibly analogous
phenomenon is the contraction of blood clots, which Budtz-Olsen (1951) has
shown is due not to the fibrin but to the platelets.
If it can be demonstrated that a population of fibroblasts can develop forces of
the magnitude of those required to cause wound contraction, the phenomenon
may have a wider significance. The embryo might in this way be provided with
a source of motive power which could be a cause of the torsions, flexures, and
transport of entire organs which occur in development.
SUMMARY
1. In order to investigate the relation of wound contraction to collagen formation, standard skin wounds were made on two groups of guinea-pigs, one group
kept on an ascorbic acid deficient diet, the other on the same diet supplemented
with ascorbic acid.
2. Ten days after wounding the amount of collagen in the repair tissue, estimated by the method of Neuman & Logan (1950), was much less in the deficient
than in the non-deficient animals.
3. Nevertheless, the amount of wound contraction which had occurred, representing a loss of 30-40 per cent, of the original wound area, was not significantly
different in the two groups.
4. Large amounts of hydroxyproline were found in the massive scabs commonly developed over the wounds of scorbutic animals.
5. It is suggested that the force that brings about wound contraction may be
developed by the connective tissue cells of the repair tissue.
ACKNOWLEDGEMENTS
We should like to express our grateful thanks to Dr. R. P. Billingham, Dr.
R. D. Harkness, Professor P. B. Medawar, and Professor J. Z. Young for their
comments on the manuscript; to Mr. H. Barker and Miss Sylvia Jenkins for their
technical help; and to the Nuffield Foundation for financial assistance.
174
M. A B E R C R O M B I E , M. H. F L I N T , A N D D. W. J A M E S
REFERENCES
ABERCROMBIE, M., FLINT, M. H., & JAMES, D. W. (1954). Collagen formation and wound con-
traction during repair of small excised wounds in the skin of rats. /. Embryol. exp. Morph. 2,
264-74.
BILLINGHAM, R. E., KROHN, P. L., & MEDAWAR, P. B. (1951a). Effect of skin homografts in rabbits.
Brit.med.J. 1,1157-63.
(1951 b). Effect of locally applied cortisone acetate on survival of skin homografts
in rabbits. Brit. med. J. 2, 1049-53.
& MEDAWAR, P. B. (1955). Contracture and intussusceptive growth in the healing of extensive
wounds in mammalian skin. /. Anat. Lond. 89, 114-23.
BOURNE, G. H. (1944). Effect of vitamin C deficiency on experimental wounds. Tensile strength
and histology. Lancet, 1, 688-92.
BRADFIELD, J. R. G., & KODICEK, E. (1951). Abnormal mucopolysaccharide and 'precollagen' in
vitamin C-deficient skin wounds. Biochem. J. 49, xvii.
BUDTZ-OLSEN, O. E. (1951). Clot Retraction. Oxford: Blackwell.
DANIELLI, J. F., FELL, H. B., & KODICEK, E. (1945). The enzymes of healing wounds. II. The
effect of different degrees of vitamin C-deficiency on the phosphatase activity in experimental
wounds in the guinea-pig. Brit. J. exp. Path. 26, 367-76.
DEVENYI, I., & HOLCZINGER, L. (1954). The morphology of wound healing under crusts. Ada
Morph. Acad. Sci. Hung. 4, 447-61.
ELSTER, S. K. (1950). Effect of ascorbic acid deficiency on collagen content of guinea-pig tissues.
/. biol. Chem. 186, 105-12.
GALLOWAY, N. M., GARRY, R. C , & HITCHIN, A. D. (1948). Ascorbic acid and epithelial regeneration. Brit. J. Nutrit. 2, 228-32.
HARTZELL, J. B., & STONE, W. E. (1942). The relationship of the concentration of ascorbic acid of
the blood to the tensile strength of wounds in animals. Surg. Gynec. Obstet. 75, 1-7.
HOFFMANN-BERLING, H. (1954). Adenosintriphosphat als Betriebsstoff von Zellbewegungen.
Biochim. Biophys. Acta, 14, 182-94.
HUBBELL, R. B., MENDEL, L. B., & WAKEMAN, A. J. (1937). A new salt mixture for use in experi-
mental diets. / . Nutrit. 14, 273-85.
HUNT, A. H. (1941). The role of vitamin C in wound healing. Brit. J. Surg. 28, 436-61.
JAMES, D. W. (1955). A connective tissue constituent in the scabs formed over healing cutaneous
wounds in rats. J. Path. Bact. 69, 33-42.
KREDEL, F. (1927). The physical relations of cells in tissue culture. Johns Hopkins Hosp. Bull.
40,216-17.
LEE, R. E., & LEE, N. Z. (1947). The peripheral vascular system and its reactions in scurvy: an
experimental study. Amer. J. Physiol. 149, 465-75.
LINDQUIST, G. (1946). The healing of skin defects. An experimental study on the white rat. Acta
chir. Scand. 94, Suppl. 107.
MAYER, E. (1933). Formbildung und Wachstum von geziichteten Zellverbanden ('Reinkulturen').
Roux Arch. EntwMech. Org. 130, 382-494.
MAZOUE, H. (1937). Etude histologique du deVeloppement de granulomes experimentaux chez des
cobayes scorbutiques. Arch. Anat. micr. 33, 129-50.
MEYER, E., & MEYER, M. B. (1944). The pathology of staphylococcus abscesses in vitamin Cdeficient guinea-pigs. Johns Hopkins Hosp. Bull. 74, 98-110.
MURRAY, P. D. F., & KODICEK, E. (1949). Bones, muscles and vitamin C. I. The effect of a partial
deficiency of vitamin C on the repair of bone and muscle in guinea-pigs. /. Anat., Lond. 83,
158-74.
NEUMAN, R. E., & LOGAN, M. A. (1950). The determination of collagen and elastin in tissues.
/. biol. Chem. 186, 549-56.
PENNEY, J. R., & BALFOUR, B. N. (1949). The effect of vitamin C on mucopolysaccharide production in wound healing. /. Path. Bact. 61, 171-8.
ROBERTSON, W. VAN B. (1950). Concentration of collagen in guinea-pig tissues in acute and prolonged scurvy. /. biol. Chem. 187, 673-7.
WOUND CONTRACTION AND COLLAGEN
15
175
ROBERTSON, W. VAN B. (1952). Influence of ascorbic acid on N incorporation into collagen in
vivo. J. biol. Chem. 197, 495-501.
TWITTY, V. C. (1949). Developmental analysis of amphibian pigmentation. Growth, Symp. 9,
133-61.
WOLBACH, S. B. (1933). Controlled formation of collagen and reticulum. A study of the source of
intercellular substance in recovery from experimental scorbutus. Amer. J. Pathol. 9, 689-99.
& BESSEY, O. A. (1942). Tissue changes in vitamin deficiencies. Physiol. Rev. 22, 233-89.
& HOWE, P. R. (1926). Intercellular substances in experimental scorbutus. Arch. Path. (Lab.
Med.) 1,1-24.
(Manuscript received 2: ix: 55)